Method of manufacturing semiconductor device having unit circuit-blocks and semiconductor devices produced thereby

Abstract

 A semiconductor device having two or more unit circuit-blocks is produced by (a) forming in a chip area on a substrate two or more circuit-blocks by repeating, a required number of times, a process comprising exposing in sequence each block section within each chip area coated with a resist film, using masks for making respective circuit-blocks, developing a resist pattern, and a subsequent process using the pattern, and (b) forming external interconnections between the circuit-blocks using a resist pattern formed over the circuit-blocks.

Description

BACKGROUND OF THE INVENTION

1. Field of the Invention

[0001] The invention relates to a multifunctional highly integrated, high capacity semiconductor device provided with unit circuit-block with separate functions, and connecting circuits for the conjunction of the circuit-blocks.

[0002] Remarkable advances in the field of semiconductor devices have been accomplished by the development of fine-forming techniques, whereby the area of use of the semiconductor devices has been rapidly enlarged, and a proportion occupied by custom integrated circuits, called application specific integrated circuits (ASICs), has increased more and more.

[0003] In the case of ASICs, in addition to the performance thereof, a shortening of a time for a delivery of products to a customer is very important. The invention provides a method of achieving a reduction of a time needed for the delivery of ASICs.

2. Description of the Related Art

[0004] Prior art ASICs are manufactured by one of the following methods: 1) a gate array method in which bulk elements such as transistors are preformed, and an interconnection thereof is then made to satisfy the user's requirements; 2) a standard cell method in which basic circuit elements are previously proposed to a user, a bulk design of a chip is made, based on the user's requirements, and the production thereof is then started; and 3) a full custom method in which the design of a chip is based, from the beginning, on the user's specification.

[0005] Each of these methods has advantages and disadvantages. For example, the gate array method provides a short delivery time, but is wasteful of unused elements, as all elements are not used when meeting a user's requirements; the full custom method eliminates this waste of elements, but requires a long delivery time, as it is necessary to design the whole chip to be manufactured as demanded by the customer.

[0006] Under the above-mentioned situation, the standard cell method is expected to dominate the ASIC manufacturing art, as it provides a balance of a delivery time and wastage of unnecessary elements.

[0007] Nevertheless, even in the standard cell method, it is important to further reduce a time needed for a delivery of products to a customer, because under present conditions, products are made by following a sequence of processes demanded by a customer, a bulk design of a chip, forming required masks, and a production of the chip, so that this method merely reduces the amount of design procedures compared with the full custom method. Moreover, since it is necessary to form all masks according to individual customer's demands, the cost of forming the masks is added to the chip price, and if the number of chips required is small, the proportion of the cost of a mask per chip is increased, and thus the chip cost becomes higher than that produced by the gate array method.

[0008] The electrode connections of a common DRAM (dynamic random access memory) are now composed of four layers of polycrystalline silicon film and two layers of aluminum film, and thus the procedure for a bulk manufacture of usual storage elements is very long, e.g., scores of to a hundred and scores of step. Currently, it is difficult to mount such large capacity storage elements on an ASIC, because of the large number of steps required by this procedure.

[0009] In the prior art, a divided exposure of a resist layer formed on a semiconductor chip is used when a single exposure operation cannot cover the whole of the chip area to be exposed to an energy ray, due to limitations in the size of a lens used. In this case, the individually exposed areas are fully merged at the boundary line therebetween, to thus form a single area having an intended function as a whole. Alternatively, the chip of the semiconductor device of the present invention is provided with two or more unit circuit-blocks or cells, each having a separate function, which are not interconnected by an underlying layer in which the unit circuit-blocks are present but are connected to each other by interconnections in an overlaying layer formed on the underlying layer. To the applicant's knowledge, semiconductor devices having such a construction have not been disclosed in the prior art.

SUMMARY OF THE INVENTION

[0010] Therefore, an object of the invention is to provide a novel method of manufacturing a semiconductor device meeting a specification required by a particular customer, with a reduced delivery time.

[0011] Another object of the invention is to provide a method capable of supplying ASIC devices at a lower cost, even if only a small number of such devices are required.

[0012] A further object of the invention is to provide a method whereby a memory with a large capacity can be mounted on an ASIC device by a lesser number of production procedures therefor.

[0013] These objects are achieved by a method of manufacturing a semiconductor device having two or more unit circuit-blocks each provided with a separate function and formed on a chip, and interconnections inter-connecting the unit circuit-blocks, the method comprising the steps of, (a) in a chip area on a semiconductor substrate, forming two or more unit circuit-blocks provided with a separate function and separated from each other, by a repetition of a process comprising, (i) sequentially exposing each block section within each chip area on the substrate coated with a resist film, using a mask for each block section, to form two or more latent images of the respective block patterns of said block sections within the chip area, (ii) simultaneously developing all of the latent images formed on the substrate, to form a resist pattern, and (iii) using the resist pattern to further process the substrate; and (b) forming external interconnections mutually and externally connecting the unit circuit-blocks on the chip area, using a resist pattern formed on the circuit-blocks by exposing a resist film coated over the substrate, using a mask for forming the resist pattern, and then developing the pattern.

[0014] The present invention also provides a semiconductor device having two or more unit circuit-blocks provided with a separate function and formed on a chip, and interconnections mutually connecting the unit circuit-blocks, wherein the unit circuit-blocks are formed on a chip, as a lower layer, between which a separating zone exists that separates the unit circuit-blocks from each other, wherein no interconnections of the separated unit circuit-blocks are included within the lower layer; interconnections mutually connecting the unit circuit-blocks being formed in an upper layer situated on the underlying layer.

BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The above and other objects and advantages of the present invention will be more fully understood from the following detailed description given with reference to the accompanying drawings, in which:

Figs. 1A to 1C schematically show various masks for making different unit circuit-blocks:

Figs. 2A to 2C schematically show masks for forming a scribing line for chips of a small, moderate, and large size, respectively;

Fig. 3 diagrammatically illustrates block sections of a chip area in which different unit circuit-blocks are formed;

Figs. 4A to 4F schematically illustrate a process for making unit circuit-blocks in a chip area;

Fig. 5A is a schematic top view of an ASIC chip provided with external interconnections between different circuit-blocks manufactured by the method of the present invention;

Fig. 5B is schematic view of the chip in Fig. 5A in a cross section taken along the line B-B;

Fig. 6 is a typical flow chart of an example of the method of the invention;

Fig. 7 shows a chip area in which five block sections are fixed;

Figs. 8A to 8F is a schematic diagram showing a process for making unit circuit-blocks as used in an embodiment of the method of the present invention; and

Fig. 9 is a schematic view of external interconnections formed on an ASIC chip according to the invention.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0016] A feature of the method of the presnt invention is that a lower layer comprising two or more unit circuit-blocks, each with a separate function, is first formed in each of chip areas on a semiconductor substrate, and interconnections mutually connecting the unit circuit-blocks are then formed as an upper layer overlaying the lower layer. Therefore, the lower layer comprises only the unit circuit-blocks and has no interconnections thereof, and the upper layer comprises only the interconnections of the unit circuit-blocks.

[0017] The term "unit circuit-blocks" as used herein defines a macro cell formed in a section of a chip and having an own function separate from those of other unit circuit-block(s) formed on the same chip. It is assumed that, when such as unit circuit-block alone is cut out of a substrate, it can be satisfactorily used as a chip having a certain function. The unit circuit-blocks are typically exemplified by gate arrays, such as random logics, adders and multipliers, memories such as various random access memories (RAMs) and read-only memories (ROMs), various central processing units (CPUs), analog integrated circuits, such as A/D converters and comparators, and other various ICs and LSIs.

[0018] In the present invention, various unit circuit-blocks are made ready for customizing an ASIC, and, prior to a customer's order, various sets of masks employed in photolithography processes for making each unit circuit-block, as well as various types of photolithography masks for forming a scribing line defining a standardized chip area, are made ready for use.

[0019] Figures 1A to 1C show various prefablicated masks 1, 1′, and 1˝ for making different blocks; a photolithography pattern is formed in the area indicated by 2, 2′, and 2˝. As can be seen, a number of masks are prepared for each unit circuit-block. For example, to make a unit circuit-block occupying a section of a chip corresponding to the configuration and dimensions of the rectangle indicated by the numeral 2 in Fig. 1A, the masks used are provided with a respective pattern formed in an area having a size similar to that of the area indicated by the numeral 2, such as for forming a diffusion, insulating or conducting layer, or contact holes, in the course of the unit block formation.

[0020] In Figs. 2A to 2C, masks 3, 3′, and 3˝ for forming a scribing line for chips of a small, moderate, and large size, respectively, are illustrated. The photolithography patterns of the scribing line, formed in the masks 3, 3′, and 3˝, are shown by the numerals 4, 4′, and 4˝, respectively.

[0021] When unit circuit-blocks to be mounted on a chip are specified by a customer, the layout of the circuit-blocks is designed in such a manner that all of the blocks are situated in a chip area of a standardized size, without leaving a surplus space, as illustrated in Fig. 3. In this drawing, the chip area 5 is surrounded by a scribing line area 6, and three circuit-blocks A, B, and C are located within the chip area 5. Based on this circuit-block layout, a wafer process for making elements in a bulk (i.e., a process from an initial oxidation of a wafer to internal interconnections in each circuit-block) is immediately started.

[0022] In a photolithography process, a resist film is coated over a wafer and baked, the wafer is realized to thereby bring a given point within a chip area 5 to the center of a mask for the block A, and the section of the chip area not covered by the mask pattern 31 (Fig. 4A) is exposed by an appropriate energy beam, to thereby form the latent image 31′ for the block A, as illustrated by the block line in Fig. 4B. Subsequently, the section of the chip area for the block B, and then the section for the block C, are exposed in this order, using mask patterns 32 and 33 for respective blocks (Figs. 4B and 4C), to thus form the latent images 32′ and 33′ for blocks B and C, as illustrated by the broken lines in Fig. 4C and 4D, in a similar way. Where an additional exposure for the scribing line is required, this is also effected by using a mask pattern 34, as illustrated in Fig. 4D.

[0023] When the exposure of the blocks and, if any, the frame-like zone for the scribing line of all the chip areas in the wafer is completed, all of the latent images are simultaneously developed to thereby form resist patterns 35, 36, 37, 38 for all of the respective blocks and scribing line, as shown in Fig. 4E. The wafer is then subjected to a subsequent process utilizing the formed resist patterns, such as an etching or ion implantation, and the resist is then removed from the wafer.

[0024] In a similar way, the combination of photolithography and the subsequent processes is repeated for the number of times required for finishing the formation of all of the unit circuit-blocks 39, 40, 41, as shown in Fig. 4F. By this stage, elements formed within the same block section are mutually connected by internal interconnections. The internal interconnections are formed only within the block section, so that the zone not occupied by the block sections is free of interconnections.

[0025] It should be noted that not all of the blocks are always subjected to all of the photolithography and subsequent processes, since the number of processes required for making a circuit-block, such as etching and film deposition, depends on the type of specific circuit-block.

[0026] In the course of this wafer process, if a pattern design of external interconnections between the circuit-blocks and the making of masks therefor are carried out, the interconnection pattern design and mask-making are completed by the time the wafer process is sent on to an external interconnection process, thereby making it possible to form the external interconnection.

[0027] The forming of the external interconnections is effect by patterning a conductive film deposited over the lower layer of circuit-blocks, as usual, using the mask for making the external interconnections.

[0028] Figure 5A is a schematic top view of an ASIC chip 7 provided with external interconnections 8 produced by the method of the present invention, and Fig. 5B is a schematic cross section view of the chip in Fig. 5A, taken along the line B-B. As shown in Fig. 5B, the circuit-blocks 9, 9′ are not interconnected in the lower layer in which they are formed, but are only interconnected in the upper layer of the external interconnections overlaying the lower layer. The external interconnections 8 are formed on an insulating film 10, over the circuit-blocks 9, 9′. The external interconnections may be formed by one or more layers. In the latter case, the upper layer of the external interconnections is composed of one or more sublayers of interconnections.

[0029] As mentioned above, a feature of the ASIC device according to the present invention is that the lower layer of the device formed on a chip provides no interconnections between the circuit-blocks, and the upper layer comprises only the external interconnections, as typically shown in Fig. 5B. Conversely, a semiconductor device according to the standard cell method of the prior art has internal interconnections within a block and external interconnections between different blocks, both of which are formed in the same layer by using a mask containing patterns of both the internal and external interconnections. An Example of internal and external interconnections of the same layer is typically shown in Fig. 7 of U.S. Patent 4,701,778, in which cells 1 and 2 represent the circuit-blocks in the present invention, and OTs and BVDD represent the internal and external interconnections, respectively.

[0030] Chips mounting circuit-blocks are finally cut out of the wafer, at the scribing line, and the chips then subjected to fabrication and testing processes, and finally, delivered to a customer in a short delivery time.

[0031] A flow chart of the above method, which is an exemplification of the present invention, is indicated in Fig. 6. As can be seen from the flow chart, according to the invention, the design of external interconnections between the circuit-blocks and the making of a mask therefor are carried out in parallel with the forming of the circuit-blocks. Accordingly, a delivery time in the present method can be reduced to 5 to 6 weeks, in contrast to that of the prior standard cell method of usually 6 to 7 weeks. If an exposure of a scribing line is required, this can be effected before or after the exposure of blocks A to C, and a simultaneous etching of the blocks and the scribing line carried out.

[0032] A more specific embodiment of the present invention is provided below, by way of example. The example is for purposes of illustration only and is not to be constructed as limiting the spirit and scope of the invention.

[0033] First, various masks having a pattern for forming a unit circuit-block are prepared, such as dynamic random access memories (DRAMs), static random access memories (SRAMs) and read-only memories of different capacities and degrees of integration, and logic gates, and these unit circuit-blocks are shown to customers.

[0034] Assuming that a customer orders an ASIC device having a complexity of elements comprising a 16 Mbit DRAM, 16 Mbit SRAM, 16 Mbit ROM, and two 1 Mgate logics, block patterns of unit circuit-blocks fitting the customer's requirements are selected, a block layout is made, and a size of the chip area 11 is subsequently fixed, taking into account of a space held as a margin for external connections between different blocks, as shown in Fig. 7. The frame-like area 12 outside the chip area is for a scribing line.

[0035] A wafer process involving an initial oxidation of a wafer, and the completion of internal interconnections within circuit-block is then initiated, using a mask having a pattern for a scribing line that matches the determined chip area and masks having a pattern matching each circuit-block.

[0036] In the wafer process, for example, when first forming an activated area within a chip of a silicon wafer, a surface of the wafer is oxidized, a silicon nitride (Si₃N₄) film is deposited on the oxidized surface, and a resist film is then coated over the wafer and baked. As shown in Fig. 8A, a scribing line pattern 51 is formed in the margin area of the chip 50 by an exposure of the area using the mask therefor, followed by a development of a latent image and an etching. Other possible patterns, such as a pattern for alignment, are also formed in this stage.

[0037] A given section of the chip area is then exposed by detecting the specified place of a DRAM, by using the alignment pattern and a mask therefor, to thereby form a latent image 52 of the block pattern of the DRAM in the resist film, as shown by a broken line in Fig. 8B. In a similar way, other latent images 53, 54, 55, 56 of the block patterns of SRAM, ROM logic 1 and logic 2 are successively formed inside the scribing line, using respective masks therefor, as illustrated in Fig. 8C to 8F.

[0038] A set of latent images 52 to 56 thus formed in the chip area, as well as the other sets of latent images formed in the other chip areas on the same wafer, are then simultaneously developed using a developer, to complete a block pattern of the resist film corresponding to a first sublayer of each unit circuit-block. An activated area pattern of the silicon nitride film is subsequently finished within each block section by etching the uncovered silicon nitride and then removing the remaining resist.

[0039] The silicon nitride film left behind is used as a mask for making a field silicon dioxide (SiO₂) film by a preferential oxidation process, and then removed.

[0040] Following the forming of the field silicon oxide film, a gate SiO₂ film is formed by a thermal oxidation process, in a similar manner.

[0041] The wafer is further processed in away similar to the forming of the activated area described above, for example, by depositing a polycrystalline silicon film over the wafer, and patterning the film to form a gate electrode and leads for a transistor. Although, in the early stage of the alignment for the forming of activated area, only the scribing line has been first exposed, developed and etched in need of the formation of the alignment mark before the divided exposure of each circuit block, a scribing line may be formed in any step of a wafer process. Accordingly, when a scribing line is formed in a photolithography step later than a first step, a simultaneous development of patterns of the scribing line and circuit-blocks may be effected.

[0042] By repeating a formation of a necessary pattern by photolithography, and a subsequent process using the formed pattern, until the internal interconnections within each unit circuit-block are completed, a production of a bulk of chips on which a lower layer comprising necessary unit circuit-blocks is mounted, is finished. At this point, all of the unit circuit-blocks are separated from each other, and no external interconnections exist between the different blocks.

[0043] Concurrently with the progress of the wafer process, an external interconnection design for mutually connecting the unit circuit-blocks formed within a chip area is carried out, and according to the design, a mask for making the external interconnections is prepared.

[0044] For making external interconnections, a phosphosilicate glass (PSG) film is deposited over the wafer finished the wafer process, contact holes are opened in the PSG film, and a conductive film of aluminum is then deposited by a sputter process. Processes up to the opening of contact holes may be carried out one by one for each unit circuit-blocks, or together for all of the blocks within the same chip area, as in a subsequent formation of external interconnections with a conductive metal.

[0045] Following the opening of contact holes and the deposition of conductive film of aluminum, and external interconnection pattern for connecting the circuit-blocks to each other in the whole chip area is exposed, developed and etched, and a surface protective insulating film is deposited, to thereby finally finish the ASIC chip. Where the upper layer of external interconnections is composed of more than one sublayer of conductive metal, the processes of opening contact holes and forming the upper sublayer interconnections are further carried out.

[0046] External interconnections are generally produced by a film of metal aluminum, but internal interconnections may be composed of a film of a metal, e.g., aluminum or a metal having a higher melting point, such as tungsten.

[0047] Figure 9 is a schematic view of external interconnections 22 formed on a PSG insulating film 21 of the ASIC chip manufactured in the above example. In this drawing, unit circuit-blocks formed beneath the insulating film 21 are indicated by broken lines, and a scribing line around the insulating film 21 is shown by the numeral 23.

[0048] In Fig. 9, the external interconnections are very roughly shown, for illustration only. In practice, a much higher number of external interconnections is made. Furthermore, the external interconnections are preferably arranged by taking spaces therebetween which represent an integer-folding of a certain unit of space, for a convenient design of external interconnections. On the other hand, internal interconnections appearing in a unit circuit-block existing in the lower layer under the external interconnection layer are generally formed with random spaces.

[0049] Memory capacities and logic scales are important factors in an ASIC device design for reducing a waste of unused elements, to satisfy customer's requirements. By forming units, such as those referred to above, as divided blocks and finally interconnecting same, various requirements from customers can be satisfied.

[0050] According to the present invention, a production of elements in unit circuit-blocks and a design of an external interconnection pattern are carried out in parallel when an ASIC device is ordered by a customer. Accordingly, the completion of the production of the elements in the blocks can be immediately followed by the external interconnection formation, to thereby allow a faster delivery of the products to the customer. Moreover, the present invention makes it possible to mount, on an ASIC device, a memory using multilayers of polycrystalline silicon films as well as a multilayered and high integrated logic which require a large number of steps for a production thereof. Thus, since the memory capacity is greatly increased by the use of recent advanced memories, and the number of logical gates is also increased, the invention makes a great contribution to an enlargement of the application of ASIC devices.

[0051] In addition, masks for which the customer must pay the expense of fabricating same are limited, and accordingly, the chip price becomes lower, even when the customer's order is for only a small number of devices, because various masks for making unit circuit-blocks in bulk can be used in common for different customers.

Claims

1. A method of manufacturing a semiconductor device having two or more unit circuit-blocks each provided with a separate function and formed on a chip, and interconnections mutually connecting the unit circuit-blocks, characterized in that the method comprises the following steps:

(a) forming in a chip area (5) on a semiconductor substrate two or more unit circuit-blocks (39, 40, 41) each provided with a separate function and separated from each other, by a repetition of a process comprising (i) exposing in sequence each block section within each chip area (5) on the substrate coated with a resist film, using a mask (31, 32, 33) for each block section, to form two or more latent images (31′, 32′, 33′) of respective block patterns of said block sections within the chip area, (ii) simultaneously developing all of the latent images (31′, 32′, 33′) formed on the substrate to form a resist pattern, and (iii) using the resist pattern to further process the substrate; and

(b) forming external interconnections (8) mutually and externally connecting the unit circuit-blocks (39, 40, 41) on the chip area, using a resist pattern formed on the circuit-blocks (39, 40, 41) by exposing a resist film coated over the substrate, using a mask for making the resist pattern, and then developing the pattern.

2. The method according to claim 1, characterized in that various sets of masks used in exposing each block sections to make each unit circuit-block (39, 40, 41) are ready for use, prior to a customer's order.

3. The method according to claim 2, characterized in that various types of mask for forming a scribing line defining a standardized chip area are ready for use, prior to a customer's order.

4. The method according to any of claims 1 to 3, characterized in that the unit circuit-blocks (39, 40, 41) are not mutually connected during the step (a).

5. The method according to any of claims 1 to 4, characterized in that the unit circuit-blocks (39, 40, 41) include one or more gate arrays, memories, central processing units, analog integrated circuits, and/or other ICs or LSIs.

6. The method according to any of claims 1 to 5, characterized in that elements formed within a unit circuit-block (39, 40, 41) are mutually connected by internal interconnections during the step (a).

7. The method according to claim 6 characterized in that the internal interconnections are made of aluminum or a metal having a higher melting point.

8. The method according to any of claims 1 to 7, characterized in that a design work of the external interconnections (8) and making of the mask used for the interconnections are carried out in parallel with the forming of the unit circuit-blocks (39, 40, 41).

9. The method according to any of claims 1 to 8, characterized in that a layer of the external interconnections is composed of one or more sublayers of interconnections.

10. The method according to claim 9, characterized in that the external interconnections (8) are made of aluminum.

11. A semiconductor device having two or more unit circuit-blocks each provided with a separate function and formed on a chip, and interconnections mutually connecting the unit circuit-blocks, characterized in that the unit circuit-blocks (9, 9′) are formed on a chip, as a lower layer, between which a separating zone exists that separates the unit circuit-blocks (9, 9′) from each other and includes no interconnections of the separated unit circuit-blocks (9, 9′) within the lower layer, and the interconnections (8) mutually connecting the unit circuit-blocks (9, 9′) are formed in an upper layer situated on the underlying layer.

12. The semiconductor device according to claim 11, characterized in that the upper layer of the interconnections mutually connecting the unit circuit-blocks (9, 9′) is composed of one or more sublayers of interconnections.

13. The semiconductor device according to claim 11 or 12, characterized in that the interconnections are composed of aluminum.

14. The semiconductor device according to any of claims 11 to 13, characterized in that the unit circuit-blocks (9, 9′) include one or more gate arrays, memories, central processing units, analog integrated circuits, and/or other ICs or LSIs.

Drawing